The Competitive Intelligence Program
Build a rolling profile of every competitor from moles and shadow surveys, then feed it to marketing.
- Difficulty
- Advanced
- Time to result
- ~ongoing to results
- Steps
- 7
- Confidence
- 90%
Patrick Campbell ran a one-person competitive-intelligence program at ProfitWell using tradecraft from his NSA analyst days. The core insight is that intel must be collected over time, not at a single point, so you build a profile rather than a snapshot. Two sources feed it: qualitative 'moles' (former employees met at conferences, friendly customers, investors weighing the space) and quantitative 'mass data' (shadow NPS and feature-preference surveys sent every one-to-three months to a competitor's identifiable customers from a neutral domain). You synthesize the stream into a timeline that reveals trends, then route the findings to the marketing team, never the product team. The output also includes 'red team docs' that pre-write your response to every existential threat.
Origin
Patrick Campbell's first job out of school was as an intelligence analyst at the NSA; he adapted target-profiling tradecraft into a competitive-intel program at his bootstrapped SaaS company ProfitWell (sold to Paddle for over $200M).
Core principles
- 01Collect over time; one data point is noise, a timeline is a profile.
- 02Two source types: qualitative moles and quantitative mass-data surveys.
- 03Give competitive intel to marketing, never to product — competitive-driven product teams build the wrong things.
- 04Use it to build a profile, not to react instantly to a single signal.
- 05Stay on the smart-collection side of the line, not the conniving side.
How to run it
- 1
Identify the competitor's customers
Harvest identifiable users from case studies, website logo walls, BuiltWith, social, and enrichment tools like Clearbit. You only need enough to sample, not the full list.
Pro tip You don't need 10,000 names — most competitors don't even have that many customers.
- 2
Stand up a neutral survey channel
Send from a generic research domain such as marketingsoftwaresearch.com, positioned as 'you've been identified as a user of X, I'd love 30 seconds on what you love and don't.'
Watch out Never spoof a look-alike domain like gethubspotfeedback.com — if it is discovered it looks terrible and crosses the ethical line.
- 3
Keep the survey to five questions
Ask the standard NPS question, most-preferred feature from a list, least-preferred from the same list, and one open comment. Respect the 30-60 second attention budget.
- 4
Re-send on a cadence to build a timeline
Send the same survey every one-to-three months. The recurring pulse turns single answers into trends you can act on.
- 5
Cultivate qualitative moles
Talk to former employees you meet at conferences, friendly customers who won't switch, and investors considering the space. Collect color, not verdicts.
- 6
Synthesize into profiles and red-team docs
Track the quantitative data over time in a spreadsheet and capture every threat-thought in a red-team doc with the pre-planned steps you'll take when it materializes.
- 7
Route findings to marketing, not product
Turn the intel into comparison pages, competitive ads, and homepage messaging. Correlationally, teams with competitive programs show better retention and lower CAC.
In the wild
As ProfitWell felt for itself that financial metrics must be ~100% accurate, its moles started reporting a rival's accuracy 'sucked.' That single synthesized insight became the company's core marketing strategy — every comparison page and competitive ad centered on accuracy, forcing rivals to react to ProfitWell rather than the reverse.
→ Accuracy became the dominant competitive message and rivals ended up copying it onto their own homepages.
Investors evaluating rivals would approach Campbell, sometimes openly asking his opinion. He showed them his synthesized slides comparing his numbers to the target's. Faced with an opponent who held that much information, at least one investor walked.
→ Campbell says he definitively killed at least one competitor's fundraising round.
Common mistakes
Giving the intel to the product team
Competitive-driven product teams build reactively and badly. The data belongs with marketing, which uses it to sharpen positioning.
Reacting to a single data point
The value is the timeline. Acting on one survey or one mole's comment leads you to over-rotate on noise.
Crossing into conniving tactics
Spoofing look-alike domains or misrepresenting yourself turns smart data collection into something that torches your reputation if exposed.
Is it for you?
Best for
Founders or marketers in a market that has turned red and competitive, who need to anticipate rival moves.
Not ideal for
Blue-ocean products with no real competitors yet, where the effort is wasted.
From the transcript
“The thing that people miss on competitive intel is like it has to be done over time. I need to understand not just a point.…”
“First, you should never show this information to your product team. It is the worst information... competitive-based marketing teams typically have much better retention and…”
From the episode
Patrick Campbell: The Billion $ Quilting Industry And Using Spy Secrets To Gain A Competitive Edge
Patrick Campbell